252345 - Memory leak in nsXMLHttpRequest.cpp (seen when using GMail)

Status: RESOLVED FIXED

(Core Graveyard :: Web Services, defect)

Description

[Philip K. Warren]

In nsXMLHttpRequest::ConvertBodyToText, a PRUnichar string is allocated with
nsMemory::Alloc, assigned into a returned buffer, but never freed.

Here is the trace from valgrind:

4244 bytes in 1 blocks are definitely lost in loss record 463 of 528
   at 0x1B9046D0: malloc (vg_replace_malloc.c:109)
   by 0x1B9E28AF: PR_Malloc (prmem.c:436)
   by 0x1BE7860B: nsMemoryImpl::Alloc(unsigned) (nsMemoryImpl.cpp:325)
   by 0x8080EE6: nsMemory::Alloc(unsigned) (nsMemory.cpp:87)
   by 0x1C50632A: nsXMLHttpRequest::ConvertBodyToText(nsAString&)
(nsXMLHttpRequest.cpp:553)
   by 0x1C506597: nsXMLHttpRequest::GetResponseText(nsAString&)
(nsXMLHttpRequest.cpp:605)
   by 0x1BE8F600: XPTC_InvokeByIndex (xptcinvoke_gcc_x86_unix.cpp:69)
   by 0x1C915CB6: XPCWrappedNative::CallMethod(XPCCallContext&,
XPCWrappedNative::CallMode) (xpcwrappednative.cpp:2027)
   by 0x1C92116F: XPCWrappedNative::GetAttribute(XPCCallContext&)
(xpcprivate.h:1887)
   by 0x1C920044: XPC_WN_GetterSetter(JSContext*, JSObject*, unsigned, long*,
long*) (xpcwrappednativejsops.cpp:1319)

Comment 1

[Philip K. Warren]

Attachment: Patch v1

Comment 2

[timeless]

in the old days one might have used adopt if the allocator fit...

Comment 3

[Philip K. Warren]

Fixed.

Checking in nsXMLHttpRequest.cpp;
/cvsroot/mozilla/extensions/xmlextras/base/src/nsXMLHttpRequest.cpp,v  <-- 
nsXMLHttpRequest.cpp
new revision: 1.115; previous revision: 1.114
done

Comment 4

[Bernard Alleysson]

It is a fixed memory leak bug, could it be considered for 1.7.2 ?

Comment 5

[Darin Fisher]

(In reply to comment #2)
> in the old days one might have used adopt if the allocator fit...

Adopt is not a method on nsAString unfortunately :-(

Comment 6

[Bernard Alleysson]

1.7.3 is done, I'd like this small memory leak to be fixed for Firefox 1.0

Comment 7

[remy cherel]

*** Bug 264901 has been marked as a duplicate of this bug. ***

Comment 8

[Benjamin Smedberg]

Comment on attachment 153835 [details] [diff] [review]
Patch v1

Whoa. Given gmail, I think we should get this on the branches. Are there any
known regressions?

Comment 9

[Asa Dotzler [:asa]]

Comment on attachment 153835 [details] [diff] [review]
Patch v1

a=asa for branches checkins.

Comment 10

[Benjamin Smedberg]

Checked in on 1.7/aviary branches.